Savannah River Site Health Effects Subcommittee (SRSHES) Meeting

Final Meeting Minutes
September 15, 2005

Back to Table of Contents

Dr. Nolan Hertel is a Professor at the Georgia Institute of Technology and served as the independent reviewer of the ATL report. CDC charged him with answering three questions to evaluate the document. First, was the methodology appropriate? Second, can an independent reviewer reproduce the results? Third, does the report contain significant errors or omissions that can significantly alter the final calculations? Dr. Hertel’s findings from his independent review of the ATL report are outlined as follows.

General Comments. ATL’s quality assurance program is solid and quite capable of tracking activities. The scenarios are reasonable based on the published literature and adequately represent the lifestyles of SRS populations. Innovations that were added for standard equations, codes and modeling approaches used in the industry are appropriate. However, repetitious text, the need to constantly switch between the report and appendices, and the absence of a table of contents and list of figures make the document difficult to read and follow.

Air Releases. Models used for air and liquid releases are appropriate and reflect standard approaches. ATL’s recommendation to analyze the air deposition of radionuclides onto ponds, reservoirs and surface runoff as a scoping calculation will not change the bounds based on the uncertainty analysis. ATL’s rationale for the fractional activity should not be used because the conservative approach of assuming plutonium to be plutonium 239 is a better reason. However, the end result is correct.

Pathways. The pathways scenario is standard. Unimportant pathways are eliminated from ATL’s analyses. Soil ingestion from the beef pathway is not analyzed, but data from the U.S. Environmental Protection Agency show that this pathway can cause exposures on the order of a fraction of a kilogram to >1 kilogram per day. The soil ingestion pathway for humans is incidental, but certain populations may be at increased risk. For example, a farmer who smokes may have more hand-to-mouth contact. The liquid pathway from swimming or accidentally ingesting water is included in ATL’s analyses, but is not fully described.

Figures selected for swimming are reasonable based on the point estimate, but whole numbers rather than decimals should be used. The method of partitioning radionuclides into separate isotopes is conservative based on the dose conversion terms. The resulting Phase III source term is reasonable. The 15 major sources of air emissions from the SRS are mapped into four virtual sources. This approach is efficient and significantly reduced the amount of data ATL had to analyze.

Meterological Data. Four five-year averages for wind data resulted in nearly average behavior. However, a determination could not be made on whether the average of four five-year averages was actually a 20-year average. Joint frequency distribution data could not be located on the CD-ROM provided to Dr. Hertel to reproduce and identify the stack that served as the source of the release. Other parameters are embedded in the text rather than in tables and are also difficult to locate.

Liquid Exposures. A reasonable technique is used to group radionuclides other than cesium, strontium and tritium that were expected to be released in liquid effluent. The Phase II analysis was reviewed and approved by the National Academy of Sciences (NAS) and is used to group radionuclides by the distribution of coefficients and estimate concentrations from other isotopes that were released into the Savannah River. Adjustments that were made to measure the frequency of cesium, strontium and tritium exposures to Lower Three Runs Creek are reasonable, but similar analyses should be performed to obtain concentrations of other isotopes in this location. However, the new analyses will not cause major changes due to the uncertainty analysis and the inclusion of the most important isotopes from actual measurements.

The analysis for liquid source terms is confusing. ATL determined that the simple model could not be used and therefore made adjustments with Phase II transport adjusted factors to create release terms for other radionuclides with no Savannah River data. The description of the simple model should be moved from the text to an appendix and revised to clearly state that the simple model could not be used. The liquid source terms are reasonable overall. The discussion on the adjustment factors is also confusing and should be clarified. For example, a determination cannot be made on whether “facility” refers to the “site.”

ATL missed the peak in terms of concentrations that were measured, calculated and used in the dose reconstruction study. ATL’s recommendation to perform additional modeling of acute releases to determine the impact of these concentrations will most likely resolve this disagreement. However, higher figures than those used in the actual data should be incorporated into new models. Data should be analyzed for the short time period of the release. Overall, the source term adequately predicts the total release over this time period. A sum of all the releases and measures will essentially provide the same result. Results of the cesium 137 concentrations in Savannah River water are appropriate.

Soil Exposures. The code ATL selected to analyze soil distribution is somewhat confusing and could not be used during the independent review to model concentrations over 39 years. The code ATL selected to analyze direct exposures solely focuses on deposition and radioactive decay and does not account for leaching. The code is limited to analyzing the buildup of concentrations in the soil for only one year because ATL’s atmospheric dispersion model only accepts constant emission rates.

Dr. Hertel’s analysis during the independent review showed that examining the buildup of soil concentrations for more than one year makes a difference. The exclusion of leaching yields a conservative result for the one-year period. The agricultural exposure includes leaching, but ignores harvesting. The result is conservative and does not account for residual from previous years. These data do not correlate with Appendix D. ATL’s position is that root uptake is not a significant contributor to dose from the ingestion of plants.

Food Chain Transport Analysis. A standard modeling approach is used. Parameter values reflect a combination of both site data and default GENII code data. These values are representative of the literature and will not drastically change results.

Dose and Risk Calculations. ATL acknowledges an error rate of <30% in external dose conversion coefficients for adults of all age groups. The 30% uncertainty range of the point estimate does not significantly change results.

Pond Data. Dr. Hertel modeled a pond with a turnover rate of one-third of the pond volume per year during the independent review, but could not locate deposition rates in the report for isotopes in closer locations to assess his equation. The turnover rate is slow and may not be realistic for private lakes or ponds.

Appendix A. The strategy of mapping 15 major release points into four virtual sources is reasonable and well within the uncertainty of other parameters. The comparison between multiple and virtual sources in all four locations is based on equal releases from the real sources and equivalent amounts from the virtual sources. The resulting uncertainty may be different, but is still well within other parameters. Problems with the GENII calculations related to concentrations and large discontinuities were corrected by slightly moving the location of sources to some towns.

ATL Results. The point estimate results are appropriate, but some tables are mislabeled. The uncertainty and sensitivity analyses were performed with methodical and solid approaches. The standard Latin hypercube sampling tool was used to conduct studies, identify parameters in the models to which results were most sensitive, and select those that would result in the highest impact. The probability distributions are based on standard assumptions for environmental parameters. A sensitivity analysis was not performed to determine differences in the amount of human consumption. Discussions of the results are solid, but somewhat lengthy.

ATL Recommendations. ATL’s recommendation to “examine the buildup of long-lived radionuclides in soil to determine if terrestrial doses significantly change” will be addressed in Dr. Hertel’s final report to CDC. ATL’s recommendation to “compare modeled concentrations in foodstuffs to monitoring data for model validation” is an extremely expensive undertaking and will most likely not change the final analysis. ATL’s recommendation to “publish papers on the study methods and results in peer-reviewed journals to obtain technical peer reviews” will add validity to the ATL report and should be considered if funding allows. ATL should perform additional modeling to determine whether surface runoff and airborne depositions into reservoirs and ponds cause significant doses. However, these results will probably not be higher than liquid effluent discharges in the Savannah River from fish concentrations.

Dr. Hertel’s overall conclusions from his independent review of the ATL report are outlined as follows. On the one hand, the report is relatively solid at this point. For the most part, the models and approaches are standard, appropriate and have generated reasonable figures with no apparent gross errors. The doses appear to be reasonable compared to background doses over a 39-year period. Implementation of different approaches will not change the 95% confidence interval of doses reported in ATL’s uncertainty analysis.

On the other hand, ATL’s outcome of generating point estimate results for the uncertainty analysis that are below median values is disappointing and will be surprising to non-statisticians. The executive summary should be revised to clarify and more strongly emphasize the point estimates, medians and means for the uncertainty analyses. However, the conclusions explain the point estimate results and the uncertainty analysis captures all possible values. The distribution is skewed to the high end of twice the radiation dose to average persons in United States from natural background sources.

Some items are difficult to locate in the report or entirely excluded, such as the pre-processor, post-processor and data to calculate the point estimate for some scenarios. An automatic setup of the subdirectories did not occur with the version of GENII provided to Dr. Hertel. The final peer review report to CDC will address the question of whether an independent reviewer can reproduce results in the ATL report and will also include a small pond model to calculate air deposition.

Dr. Lee noted that the scenarios in the ATL report represent the highest dose to an infant and beef ingestion as the highest pathway. She asked Dr. Hertel to analyze these data and include comments in his final report on whether ATL’s approach is sound. She also pointed out a significant gap. The ATL report does not contain sufficient data on the pre-processor to identify specific figures and uncertainty ranges from the Phase II report that served as the basis to calculate doses in the Phase III report. Persons with expertise in dose reconstruction are not able to easily run ATL’s codes and reproduce the figures.

Mr. Wood summarized comments on the ATL report in response to SRSHES’s previous action item. Many comments reiterated Dr. Lee’s concerns that ATL’s approach to calculate doses is unclear and data are missing from the report to reproduce tables. Requests were made to calculate downstream river doses. Concerns were expressed that doses were not calculated for specific populations in certain areas, such as water doses to Jasper, South Carolina residents. For the final report, CDC will instruct ATL to clarify and emphasize its methodology of selecting representative scenarios to ensure the public understands that doses were not calculated for each SRS community.

Mr. Wood added that several comments asked CDC to conduct more studies in the future to address ATL’s conclusions and recommendations outlined in Chapter 13. CDC’s response to these requests is that DOE will only allocate $2.7 million in FY’06 for health-related research at its facilities. This funding will be shared by NCEH/ATSDR and NIOSH and will cover all DOE sites. As a result, CDC will be extremely reluctant to support additional calculations, computer models or data collection unless the outcomes are reasonably expected to demonstrate significant health exposures or larger doses to persons. DOE’s FY’06 funding for health-related research at its facilities will not be sufficient for CDC to merely improve the ATL report.

Mr. Wood reported that with the exception of Dr. Hertel’s draft independent review, comments from the public and all other sources submitted to CDC to date were compiled and provided to ATL. ATL was also given SRSHES’s editorial and technical comments presented during the previous meeting.